paweld/fpc
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/fpc/source.git synced 2025-04-06 14:28:01 +02:00
Code Issues Packages Projects Releases Wiki Activity
71,844 Commits 100 Branches 75 Tags 315 MiB
main
Commit Graph

130 Commits

Author SHA1 Message Date
Sven/Sarah Barth
e94d02a067 * with all existing RTLs switched over to the atomic intrinsics, the define FPC_SYSTEM_INTERLOCKED_USE_INTRIN can be removed again 2024-12-12 22:05:20 +01:00
Sven/Sarah Barth
a815beea2c * switch M68k RTL to provide atomic intrinsic helpers instead of Interlocked* functions 2024-12-12 22:05:16 +01:00
Rika Ichinose
ce5a632432 Remove the rest of the VER3_0 conditionals. 2024-09-01 17:13:39 +00:00
Jonas Maebe
0758aa1143 FPU exception mask: generlised system unit interface 2022-10-17 19:43:01 +00:00
Pierre Muller
870d612e15 Fix compilation of m68k-linux rtl with -dFPC_USE_LIBC 2022-06-08 00:20:57 +02:00
Karoly Balogh
ecef61e9c5 * m68k: fix an instruction which doesn't exist on ColdFire 2022-02-12 17:32:59 +01:00
Karoly Balogh
12ef9a4a22 * m68k: some CPUs don't have TAS either (for example some CF) 2022-02-11 01:53:19 +01:00
Karoly Balogh
3c5d6e7224 * m68k: implement basic support for interlocked functions using a TAS based global spinlock on CPUs which don't have the CAS instruction 2022-02-08 04:42:30 +01:00
pierre
3362abb30c * Set softfloat_rounding_mode indise SetRoundMode function for all CPUs. 
* SetRoundMode returns previous rounding mode value for all CPUs.

git-svn-id: trunk@48018 -
 2021-01-03 21:44:18 +00:00
pierre
90d2982977 Wipe out any pre-existing exception on SetExceptionMask call for m68k cpu 
git-svn-id: trunk@43908 -
 2020-01-10 22:56:03 +00:00
pierre
d8298c9f14 Try to correctly handle m68k exceptions bits inside SetExceptionMask function 
git-svn-id: trunk@43893 -
 2020-01-08 23:08:30 +00:00
Károly Balogh
3a78ff1ee4 m68k: on setjmp/longjmp also save/restore the FPCR register when FPU support is enabled 
git-svn-id: trunk@38522 -
 2018-03-13 23:44:54 +00:00
svenbarth
9240e0c2e6 * fix 32-bit * 32-bit => 64-bit for Coldfire 
git-svn-id: trunk@38060 -
 2018-01-27 14:26:18 +00:00
svenbarth
8522474457 * correctly allocate the space for the 5 registers that are stored 
git-svn-id: trunk@38054 -
 2018-01-26 21:48:07 +00:00
svenbarth
bd57af5b19 * as Coldfire does not support ADD.W save the (original) lower word and restore it after D0 is swapped back (I hope the algorithm still works then O.o) 
git-svn-id: trunk@38044 -
 2018-01-25 20:02:16 +00:00
svenbarth
b7cdec3a2a * Coldfire does not support EXG, so do it the old fashioned way (but before restoring the stored registers) 
git-svn-id: trunk@38043 -
 2018-01-25 19:59:50 +00:00
svenbarth
9822374680 * Coldfire does not support MOVEM together with pre-decrement or post-increment, so abuse TST for that which is only a single instruction word 
git-svn-id: trunk@38042 -
 2018-01-25 19:58:59 +00:00
svenbarth
00141749d5 * Coldfire does not support sub.w, so use sub.l instead for that CPU 
git-svn-id: trunk@38041 -
 2018-01-25 19:56:28 +00:00
Károly Balogh
c054117870 m68k: an assembly optimized longint to int64 implementation. removed the sarint64 helper from here 
git-svn-id: trunk@37825 -
 2017-12-26 22:41:06 +00:00
Károly Balogh
8326543918 m68k: apparently, the right place for the sar qword helper is still in the CPU include, not the 64bit helper include. this fixes linux m68k build. 
git-svn-id: trunk@37820 -
 2017-12-26 18:05:34 +00:00
Károly Balogh
7659b347cd m68k: assembly optimized helper for the SarInt64 intrinsic 
git-svn-id: trunk@37819 -
 2017-12-26 16:34:01 +00:00
florian
ff86c80176 * cpu specific header file for the system unit: cpuh.inc, moved several declarations into it 
git-svn-id: trunk@37542 -
 2017-11-01 16:33:30 +00:00
Károly Balogh
e1501c86bf m68k: fix no FPU support 
git-svn-id: trunk@37227 -
 2017-09-16 18:13:57 +00:00
Károly Balogh
f4a8a29aea m68k: try to restore compilability with a 3.0 compiler 
git-svn-id: trunk@37026 -
 2017-08-22 09:05:20 +00:00
svenbarth
2c628e983d * fix define so that m68k for Coldfire builds correctly again 
git-svn-id: trunk@36688 -
 2017-07-08 20:29:25 +00:00
Károly Balogh
17e85c4a09 m68k: rewritten FillChar from scratch. does aligned word and dword writes on large block fills. depending on the host CPU it's up to 5x faster on medium and large block fills 
git-svn-id: trunk@36631 -
 2017-07-03 19:22:51 +00:00
Károly Balogh
2d8313165d m68k: support FFU exception settings, also reworked FPU initalization 
git-svn-id: trunk@36618 -
 2017-06-30 02:38:14 +00:00
Károly Balogh
30176f3116 m68k: enable inlined get_frame for m68k 
git-svn-id: trunk@36577 -
 2017-06-22 17:43:24 +00:00
Károly Balogh
238f423c71 m68k: implement the interlocked family of functions, to be used on Linux and BSDs 
git-svn-id: trunk@36557 -
 2017-06-21 01:02:49 +00:00
Károly Balogh
50812415d8 m68k: add a modified version of the generic fpc_round_real, which takes some m68k FPU oddities into account 
git-svn-id: trunk@36531 -
 2017-06-19 01:12:00 +00:00
Károly Balogh
f1ee65610d m68k: added support for setting the rounding mode on hardware FPUs 
git-svn-id: trunk@36523 -
 2017-06-18 19:06:34 +00:00
Károly Balogh
a635f435b9 m68k: replaced fpc_longword_to_double function with a working and branchless version. fixes tcnvint4 test among others with FPU enabled. 
git-svn-id: trunk@36503 -
 2017-06-15 02:22:49 +00:00
Károly Balogh
35ddac66fc m68k: fix frame pointer register on atari, and a few other non-amiga systems 
git-svn-id: trunk@35204 -
 2016-12-27 17:09:06 +00:00
Károly Balogh
7d2360eb0c m68k: Test68000/Test68881 support boilerplate and special support for Amiga 
git-svn-id: trunk@35194 -
 2016-12-24 21:11:06 +00:00
Károly Balogh
6a856f109e m68k: made setjmp and longjmp code work when register calling convention is the default 
git-svn-id: trunk@34826 -
 2016-11-06 19:10:47 +00:00
marcus
e54fdc3853 m68k: fixed compilation without FPU after 34785 
git-svn-id: trunk@34803 -
 2016-11-06 08:42:41 +00:00
Károly Balogh
7293bb7fdb m68k: reworked setjmp/longjmp to also save the FPU registers (when compiled with HW FPU support), and to only save the nonvolatile registers 
git-svn-id: trunk@34785 -
 2016-11-05 22:02:50 +00:00
Károly Balogh
391a79aebf m68k: a bunch of tweaks and improvements to assembly functions. higher move() speed on coldfire, disabled fillword implementation on CPUs not supporting unaligned access 
git-svn-id: trunk@33805 -
 2016-05-25 22:44:18 +00:00
Károly Balogh
846dd6b57b m68k: use the Amiga-specific atomic ops on Amiga 
git-svn-id: trunk@30903 -
 2015-05-24 21:44:51 +00:00
Károly Balogh
49f3473735 fixed the define around the FPU code 
git-svn-id: trunk@30695 -
 2015-04-21 17:05:43 +00:00
Károly Balogh
7758ad9313 m68k: clear the FPU Control Register at program start. This fixes some testcode randomly failing on AmigaOS, where the FPCR state is inherited from the parent task. Also added the FPCR layout and rounding modes info, based on the Motorola 68k PRMm68k.inc 
git-svn-id: trunk@30060 -
 2015-03-02 00:51:23 +00:00
Károly Balogh
f2c7a3efbf m68k: some assembler SwapEndian() implementations I did a while back. disabled for now, because they weren't extensively tested and benchmarked 
git-svn-id: trunk@29380 -
 2015-01-04 01:39:49 +00:00
sergei
38d5be4182 + Comment about behavior of DBRA instruction 
git-svn-id: trunk@29321 -
 2014-12-25 09:32:10 +00:00
sergei
6bcaa88c3a * m68k: "or" without suffix is not accepted, must be "or.l" 
git-svn-id: trunk@28184 -
 2014-07-08 18:48:57 +00:00
sergei
cfe13734e0 * m68k: Fixed damage caused by r27573 and r28177: 
* longjmp: restored loading of function result (d0) which got removed in r28177.
  * While at the point, added test for zero, because longjmp must never return zero values.
  * Assembler reader: 'fp' refers to frame pointer, not the stack pointer.

git-svn-id: trunk@28183 -
 2014-07-08 15:04:54 +00:00
Károly Balogh
f98cd634fb m68k: reworked/cleaned setjmp/longjmp code to use the movem instruction with register lists 
git-svn-id: trunk@28177 -
 2014-07-06 13:18:56 +00:00
sergei
e232dd5066 + m68k: default implementation of fpc_cpuinit, provides correct masking of floating-point exceptions and fixes test/testv8.pp. 
git-svn-id: trunk@28128 -
 2014-07-02 16:23:40 +00:00
sergei
23a20f0e50 * m68k: fixed setjmp/longjmp: 
* Data registers have to be saved/restored using 'move.l', because 'move' without suffix assumes 16-bit size.
  * setjmp must behave as a regular function w.r.t. removing its argument from the stack.

git-svn-id: trunk@28103 -
 2014-07-01 11:46:10 +00:00
sergei
acd3ea8750 * m68k: Fixed parameter passing to conform to ABI: 
* records are passed by value
  * records with size of 1,2 or 4 are returned in registers
  * parameters with size<4 are justified on the stack according to big-endian target

Now everything except floating-point parameters is compatible with C code compiled with "-malign-int -mrtd".
Compatibility with "-mno-align-int" is achievable by changing target_info.maxCrecordalign to 2, but doings so causes a lot more troubles because RTL (incorrectly) assumes that records declared with {$PACKRECORDS C} are aligned to pointer size.

+ Reuse parameter locations. Since everything is passed on stack, it reduces code size quite a bit.
- tm68kparamanager.getintparaloc removed, generic implementation has been tested and works as expected.

git-svn-id: trunk@28083 -
 2014-06-27 06:58:39 +00:00
sergei
217ab9879c * Enabled internal handling of Abs(longint) for all targets. It has been implemented in cross-platform way ages ago (see tcginlinenode.second_abs_long), but not enabled on MIPS,SPARC and m68k. 
- RTL: removed MIPS,SPARC and m68k-specific implementations of Abs(longint), and marked the generic one as  required for bootstrapping purposes only.

git-svn-id: trunk@27857 -
 2014-06-05 10:35:51 +00:00